US 20080244611A1 (19) United States (12) Patent Application Publication (10) Pub. No.: US 2008/0244611 A1 Doyle et al. (43) Pub. Date: (54) PRODUCT, METHOD AND SYSTEM FOR (21) Appl. No.: 11/692,137 IMPROVED COMPUTER DATA PROCESSING CAPACITY PLANNING USING (22) Filed; Man 27, 2007 DEPENDENCY RELATIONSHIPS FROM A CONFIGURATION MANAGEMENT.... DATABASE Publication Classi?cation (51) Int. Cl. (75) Inventors: Ronald Patrick Doyle, Raleigh, G06F 9/46 (2006.01) NC (U S); David Louis Kaminsky, Chapel Hill, NC (Us) (52) us. Cl...... 718/105 Correspondence Address: LEE LAW, PLLC (57) ABSTRACT IBM CUSTOMER NUMBER P_()_ BOX 189 The invention discloses a computer data processing capacity PITTSBORO, NC 27312 (Us) planning system that utilizes known Workload planning infor mation along With hardware and/or software con?guration (73) Assignee; INTERNATIONAL BUSINESS information from the actual operating environment to accu MACHINES CORPORATION, rately estimate the production system capacity available for Armonk, NY (US) use in carrying out one or more processing task(s). configuration management database (CMDB) 20 dependency information 120 neighborhood information 12] capacity planning algorithm 100 requirements table 110 Workload modeling pro gram 10
Patent Application Publication US 2008/0244611 A1 configuration management database (CMDB) 20 dependency information 120 neighborhood information 121 capacity planning algorithm 100 requirements table 110 I Workload modeling program 10 FIG. 1
PRODUCT, METHOD AND SYSTEM FOR IMPROVED COMPUTER DATA PROCESSING CAPACITY PLANNING USING DEPENDENCY RELATIONSHIPS FROM A CONFIGURATION MANAGEMENT DATABASE TECHNICAL FIELD [0001] This invention relates to computer system data pro cessing capacity planning utilizing information stored in a con?guration management database (CMDB). BACKGROUND [0002] A con?guration management database (CMDB) provides a mechanism for specifying and storing system con?guration artifacts (i.e., computer hardware and/or soft Ware processing resources) and dependencies (i.e., the data and/ or parameters used in de?ning the relationships) among those artifacts for accomplishing a given set of processing task(s). The artifacts include a speci?cation of the hardware and software components (identifying computer processors, memory, peripherals, clients/ servers, routers, application and/or operating system software, databases and transaction tiers, etc.) Which make up the necessary system processing infrastructure, While dependency information includes con?guration data de?ning the physical and/or functional and/ or informational relationships between these processing resources used for meeting business or other production goals (for example, this information can be populated in a CMDB using program tools such as Collation Con?gnia). [0003] Known techniques for system capacity planning involve estimating the expected Workload (in terms of extent and rate of data input and/or output) While also accounting for any limitations and resource requirements of the software (and of the physical processing and/or memory components) used for a given application in order to determine the neces sary resources to allocate in meeting production require ments, such as planning the server capacity needed to accom modate the peak and normal loads of a data processing task. One of the drawbacks of traditional capacity planning is that these estimates are made in the laboratory When designing a system for operation but do not account for aspects of the actual production environment, Whereas the invention enables capacity planning to be more accurately performed by leveraging information available about the processing environment that exists When the system is put into operation. SUMMARY OF THE INVENTION [0004] The invention provides a product, method and sys tem for improved computer system data processing capacity planning using dependency relationships from a con?gura tion man Speci?cally, a computer data pro cessing capacity planning system is disclosed that utilizes known Workload planning information along With hardware and/or software con?guration information from the actual operating environment to accurately estimate the production system capacity available for use in carrying out one or more processing task(s). [0005] The invention discloses a computer system data pro cessing capacity planning algorithm that utilizes traditional input(s) to a Workload modeling program as Well as neigh borhood information from a con?guration management database (CMDB) identifying parameters for any other soft Ware process(es) and/or con?guration(s) running on available hardware (along With the planned task(s)) in formulating an estimate of the expected system data processing Workload at any (or all) point(s) in time. The output of the capacity plan ning algorithm program is a speci?cation of con?guration parameters used for allocating performance of required soft Ware processing functions to assigned hardware. [0006] In operation, the capacity planning program algo rithm can notify system user(s) of con?guration change events having a negative (or positive) impact on previously planned tasks (through diminishment or enhancement of sys tem processing resources or services) as the CMBD is auto matically or manually updated With changes to the system hardware and/or software con?guration (arising Within the framework of a Work?oW controlled processing environ ment) Which allows a user to approve (or reject) a con?gura tion change depending on its impact to system capacity. [0007] It is therefore an object of the present invention to provide a product, method and system for improved computer system data processing capacity planning using dependency relationship information stored in a con?guration manage ment database (CMDB) to accurately estimate the production system capacity available for use in carrying out one or more processing tasks. [0008] It is another object of the present invention to pro vide a computer system data processing capacity planning algorithm that utilizes known Workload modeling program inputs as Well as information from a con?guration manage ment database (CMDB) identifying parameters for any other software processes or con?gurations running on available hardware in addition to the planned tasks in formulating an estimate of the expected system data processing capacity at any point in time. [0009] The subject matter Which is regarded as the inven tion is particularly pointed out and distinctly claimed in the concluding portion of the speci?cation. The invention, how ever, together With further objects and advantages thereof, may best be understood by reference to the following descrip tion taken in conjunction With the accompanying drawings. BRIEF DESCRIPTION OF THE DETAILED DRAWINGS [0010] FIG. 1 illustrates the components of a computer system data processing capacity planning program according to the invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS [0011] Many computer databases and other applications (such as DB2 or WebSphere ) are used for matching hard Ware and/or software capacity to the estimated processing Workload. FIG. 1 illustrates a system for implementing a computer data processing capacity planning algorithm 100 to extend the scope of software and hardware dependency information 120 used in such traditional Workload modeling capacity planning systems 10 relating to planned processing tasks, by including parameters 121 acquired from a con?gu ration management database (CMDB) 20 relating to other (possibly independent) operating software (i.e., size of memory buffer pools, number of active program process threads, speci?cation of connections to database back end servers, etc.) to determine the system processing resources needed to meet all production requirements. The
algorithm can also use other system neighborhood informa tion 121 gathered from the CMDB 20 (such as reduced processing capacity arising from hardware and/or software failures or components not con?gured for use) in making this calculation. [0012] The capacity planning program algorithm 100 pref erably has the following input parameters 120: (l) estimated Workload (in terms of amount and/or rate of data input and/or output) required to accomplish a planned processing task (preferably obtained using a program such as IBM Opera capacity planner); (2) calculated system processing response times measured from known data input arrival rates (prefer ably under a variant of Littles LaW); (3) known or measured hardware processing capacity (i.e., computer processor/ CPU speed, memory size/speed, data storage disk throughput, etc.); (4) formulas (that can be based on performance research) for calculating the effect of varying processor speed and/or memory allocation on Workload output; and (5) neighborhood information 121 consisting of other system software resource usage (either gathered directly and/or learned from application pro?les and/or retrieved from per formance measurements of the software con?guration) stored in the CMDB 20. [0013] Using these inputs, the capacity planning program algorithm performs an analysis by generating a require ments table 110 of system hardware and/or software pro cessing resources needed to meet speci?ed production per formance levels, Which is then used in a constraints-based planning approach of matching (or?tting ) the needed software applications to the hardware systems utilized in processing them based on the required (and available) system resource types and their capabilities (performed in a manner similar to solving a traditional bin-packing problem using multiple processing resources as inputs). In cases Where a solution is intractable, restrictions are imposed involving resource prioritization (for example CPU allocation may be given more importance than memory allocation and/ or data storage disk speed, etc.) to provide a locally optimum solu tion. Once the planning stage is complete, the algorithm performs a placement/validation phase to con?rm that hardware and/or software allocation constraints imposed by the plan are met, Which assures that adequate processing capacity exists and that system resources Will be reliably available for performing the designated processing tasks. [0014] The output of the capacity planning algorithm pro gram is an assignment of software resources to processing hardware, along With con?guration parameters used for allo cating performance of the software processing functions to the assigned hardware. In the example of an HTTP server, the con?guration set may include the amount of physical memory dedicated for cache space, or the amount of CPU processing time assigned to a particular logical partition (LPAR) or hypervisor slice hosting the database or server application. The algorithm can be instituted at the beginning (i.e., the roll-out phase of system deployment) or it can be invoked based on event noti?cation from the CMDB of con?guration changes (particularly in the case of software not included in the original plan). In the case of initial capacity planning, a placement stage can also be instituted to bin pack system software components to?t new hardware resources that Were not included in the previous plan. [0015] While certain preferred features of the invention have been shown by Way of illustration, many modi?cations and changes can be made that fall Within the true spirit of the invention as embodied in the following claims, Which are to be interpreted as broadly as the law permits to cover the full scope of the invention, including all equivalents thereto. What is claimed is: 1. A computer system for data processing capacity plan ning using hardware and software resource dependency infor mation stored in a con?guration management database and comprised of at least the following components containing program instructions executed to estimate available system capacity for carrying out one or more processing tasks: (a). a Workload modeling program comprised of instruc tions for using one or more inputs from a con?guration management database to formulate an estimate of sys tem data processing capacity at a point in time; and (b). a con?guration management database comprised of one or more parameters relating to any software con?gu rations running on available hardware as part of or in addition to one or more planned processes; 2. The computer system of claim 1 Wherein the con?gura tion management database is updated With changes to system hardware or software con?guration to allow a user to approve or reject a con?guration change depending on impact to sys tem capacity. 3. The computer system of claim 1 Wherein the Workload modeling program executes a capacity planning algorithm comprised of one or more of the following inputs: 4. The computer system of claim 3 Wherein resource pri ority restrictions are imposed or system software is assigned to newly available hardware to provide optimum system data processing capacity. 5. A method of using a computer system for data process ing capacity planning With hardware and software resource dependency information stored in a con?guration manage ment database and comprised of at least the following steps carried out by the following components containing program instructions executed to estimate available system capacity for carrying out one or more processing tasks: (a). instructing a Workload modeling program to use one or more inputs from a con?guration management database to formulate an estimate of system data processing capacity at a point in time; and (b). programming a con?guration management database comprised of one or more parameters relating to any software con?gurations running on available hardware as part of or in addition to one or more planned pro cesses; 6. The method of claim 5 Wherein the con?guration man agement database is updated With changes to system hard
Ware or software con?guration to allow a user to approve or reject a con?guration change depending on impact to system capacity. 7. The method of claim 5 Wherein the Workload modeling program executes a capacity planning algorithm comprised of one or more of the following inputs: 8. The method of claim 7 Wherein resource priority restric tions are imposed or system software is assigned to newly available hardware to provide optimum system data process ing capacity. 9. A computer product used With a computer system for data processing capacity planning With hardware and soft Ware resource dependency information stored in a con?gura tion management database and comprised of a computer readable storage medium containing program instructions executed by at least the following components to estimate available system capacity for carrying out one or more pro cessing tasks: (a). a Workload modeling program comprised of instruc tions for using one or more inputs from a con?guration management database to formulate an estimate of sys tem data processing capacity at a point in time; and (b). a con?guration management database comprised of one or more parameters relating to any software con?gu rations running on available hardware as part of or in addition to one or more planned processes; 10. The computer product of claim 9 Wherein the con?gu ration management database is updated With changes to sys tem hardware or software con?guration to allow a user to approve or reject a con?guration change depending on impact to system capacity. 11. The computer product of claim 9 Wherein the Workload modeling program executes a capacity planning algorithm comprised of one or more of the following inputs: 12. The computer product of claim 11 Wherein resource priority restrictions are imposed or system software is assigned to newly available hardware to provide optimum system data processing capacity. * * * * *